Verisk Marketing Solutions

Business Challenge

In the last three years, the Verisk Marketing Solutions (VMS) business unit was formed through the acquisition of Jornaya, a leader in real-time behavioral data and marketing intelligence, and Infutor, a major consumer identity resolution and intelligence provider. As a result, the VMS data science team had a slew of challenges, from integrating all new data-driven products to creating new ones utilizing novel data sources, economies of scale, big data and cutting-edge tools.

Analytics Solution

Through the acquisitions, VMS gained access to novel sources of consumer intelligence and an unprecedented scale of data, allowing for supercharged identity resolution, data enrichment and in-market signal products, all of which are based on complex data science and machine learning. The data science team developed an enhanced deterministic identity resolution algorithm using graph machine learning and big data to ensure that clients get the most relevant information about their consumers. Additionally, the data science team rebuilt the event classification machine learning pipeline responsible for categorizing all incoming consumer shopping signals using Amazon EMR with Apache Spark, Amazon Web Services and hundreds of millions of incoming records daily.

Impact

The synergy of previously disparate data sources and advanced data science practices resulted in a significantly positive impact across all existing products. VMS’ identity resolution capabilities have seen increased match rates with higher accuracy. The company’s event classification models are consistently categorizing hundreds of millions of consumer shopping signals with high precision daily and are resistant to data drift. The data science team at VMS has also developed new data products to market in collaboration with strategy, product and sales teams by bringing data from previously separate companies together.